In wzthu/ATACpipe: An Easy-to-use Systematic pipeline for ATACseq data analysis

knitr::opts_chunk$set(echo = TRUE)

load("Report.Rdata")

Summary Table

Sequence files below are set as inputs of the pipeline.

filelist

Summerized infomation on sequence files has been shown showed below. You can see details in later sections

knitr::kable(x = wholesummary)

A nucleosome free region (NFR) must be present. A mononucleosome peak must be present in the fragment length distribution. These are reads that span a single nucleosome, so they are longer than 147 bp but shorter than 147*2 bp. Good ATAC-seq datasets have reads that span nucleosomes (which allows for calling nucleosome positions in addition to open regions of chromatin).

library(ggplot2)
readsCounts<-getReportVal(atacProcs$fragLenDistr,"readsCounts")
ggplot(readsCounts[1:1000,], aes(length,counts))+geom_path(color="Red")+xlab("Fragment length (bp)")+ylab("Read counts") + theme_bw() + theme(panel.grid =element_blank()) + geom_vline(xintercept = c(147,294), linetype=2) + annotate("text", x = 147, y = max(readsCounts[1:1000,2]),label="147bp") + annotate("text", x = 147*2, y = max(readsCounts[1:1000,2]),label="147bp*2")+ labs(title = "Fragment Length Distribution") + theme(plot.title = element_text(hjust = 0.5))

Sequence Statistics

FastQC

Quality control for the sequence data

QC_path <- getReportVal(atacProcs$atacQC,"pdf")

Click to Visit Report

Remove adapter

The adapter sequence are shown below. For paired end reads, if adapters were not setted, the adapters below are identified by AdapterRemoval.

knitr::kable(getReportVal(atacProcs$removeAdapter,"adapters"))

The statistic of adapter removing are show below.

knitr::kable(getReportVal(atacProcs$removeAdapter,"statistics"))

For detail, you can visit Website of AdapterRemoval on Github.

Reads Alignment Statistics

Bowtie2 alignment log

getReportVal(atacProcs$bowtie2Mapping,"detail")

Library complexity

knitr::kable(x = getReportVal(atacProcs$libComplexQC,"report"))

The annotation you can see in section 1.

Filtering statistics

knitr::kable(x = filtstat)

Fragment size distribution

library(ggplot2)
load("Report.Rdata")
readsCounts<-getReportVal(atacProcs$fragLenDistr,"readsCounts")
ggplot(readsCounts[1:1000,], aes(length,counts))+geom_path(color="Red")+xlab("Fragment length (bp)")+ylab("Read counts") + theme_bw() + theme(panel.grid =element_blank()) + labs(title = "Fragment Length Distribution") + theme(plot.title = element_text(hjust = 0.5))

library(stats)
strength<-Mod(fft(readsCounts$counts))/length(readsCounts$counts)
periodx<-length(readsCounts$counts)/(1:(length(readsCounts$counts)-1))
strength<-strength[2:length(strength)]

rs1<-as.data.frame(cbind(periodx[periodx<20&periodx>2],strength[periodx<20&periodx>2],0))
rs2<-as.data.frame(cbind(periodx[periodx<400&periodx>2],strength[periodx<400&periodx>2],1))
rs<-rbind(rs1,rs2)
colnames(rs)<-c("period","strength","check")

g1<-ggplot(rs[rs["check"]==0,]) + geom_vline(xintercept = 10.4, linetype=2)+ geom_line(aes(x=period,y=strength),color="Red")+ theme_bw() + theme(panel.grid =element_blank()) + annotate("text", x = 10.4, y = max(rs[rs["check"]==0,2]), label = "10.4bp") +xlab("period") + ylab("strength")+ labs(title = "the Pitch of the DNA Helix") + theme(plot.title = element_text(hjust = 0.5))

g2<-ggplot(rs[rs["check"]==1,]) + geom_vline(xintercept = 186, linetype=2)+ geom_line(aes(x=period,y=strength),color="Red")+ theme_bw() + theme(panel.grid =element_blank()) + annotate("text", x = 186, y = max(rs[rs["check"]==1,2]), label = "186bp") +xlab("period") + ylab("strength")+ labs(title = "Nucleosome") + theme(plot.title = element_text(hjust = 0.5))
library(gridExtra)
grid.arrange(g1, g2, ncol=2)

TSS enrichment

The nucleosome free reads (<100bp) and monnucleosome span reads (180~247bp) enrichment around transcription starting site (TSS) are shown below.

library(ggplot2)
load("Report.Rdata")
df<-getReportVal(atacProcs$tssqc100,"tss")
g1<-ggplot(df,aes(pos,counts))+geom_line()+ geom_vline(xintercept = 0, linetype=2)+xlab("upstream<-TSS->downstream")+ylab("reads count")+theme_bw() + theme(panel.grid =element_blank()) + labs(title = "Nucleosome Free Reads") + theme(plot.title = element_text(hjust = 0.5))
df<-getReportVal(atacProcs$tssqc180_247,"tss")
g2<-ggplot(df,aes(pos,counts))+geom_line()+ geom_vline(xintercept = 0, linetype=2)+xlab("upstream<-TSS->downstream")+ylab("reads count")+theme_bw() + theme(panel.grid =element_blank())  + labs(title = "Monnucleosome Span Reads") + theme(plot.title = element_text(hjust = 0.5))
grid.arrange(g1, g2, ncol=2)

Peak Statistics

Blacklist ratio

knitr::kable(x = getReportVal(atacProcs$blacklistQC,"report"))

DHS ratio

knitr::kable(x = getReportVal(atacProcs$DHSQC,"report"))

Fraction of reads in peaks (FRiP)

knitr::kable(x = getReportVal(atacProcs$fripQC,"report"))

Peak Annotation

library(ChIPseeker)
peakanno <- getReportVal(atacProcs$Peakanno,"annoOutput.rds")
plotAnnoPie(x = peakanno)

Annotation of items in table

For single end sequencing data, esATAC will counts reads number.

For paired end sequencing data, esATAC will counts read pairs or fragment number.

• Sequence files type is the type of sequencing data: single end data and paired end data. If paired end reads are stored in one file interleavedly rather than two files, "it is call "interleaved.

• Original total reads is the sample's raw total reads (pairs) number.

• Reads after adapter removing (ratio) is the reads (pairs) number after adapter removing and the percentage of retained reads in original total reads. The larger value shows the better quality.

• Total mapped reads (ratio)
  is the reads (pairs) number mapped to reference genome and the percentage of mapped reads in original total reads (alignment rate). ENCODE recommend that the alignment rate, or percentage of mapped reads, should be greater than 95%, though values >80% may be acceptable.

• Unique locations mapped uniquely
  is the number of distinct uniquely mapping reads (i.e. after removing duplicates).

• Non-Redundant Fraction (NRF) is the value of: Unique locations mapped uniquely (the number of positions in the genome that uniquely mappable reads map to) / the total number of uniquely mappable reads.

ENCODE recommend NRF>0.9 for ATAC-seq data.

$NRF$ value range |Complexity :------------------:|:---------: $NRF<0.7$ |Concerning $0.7\le NRF \le 0.9$|Acceptable $NRF>0.7$ |Ideal

• Locations with only 1 reads mapping uniquely is the number of genomic locations where exactly one read maps uniquely.

• Locations with only 2 reads mapping uniquely is the number of genomic locations where two reads map uniquely.

• PCR Bottlenecking Coefficients 1 (PBC1) is the value of: Locations with only 1 reads mapping uniquely / Unique locations mapped uniquely. ENCODE recommend that PBC1>0.9 for ATAC-seq data.

$PBC1$ value range |Bottlenecking level :-------------------:|:---------: $PBC1<0.7$ |Severe $0.7\le PBC1 \le 0.9$|Moderate $PBC1>0.7$ |None

• PCR Bottlenecking Coefficients 2 (PBC2) is the value of: Locations with only 1 reads mapping uniquely / Locations with only 2 reads mapping uniquely. ENCODE recommend that PBC2>3 for ATAC-seq data.

$PBC2$ value range |Bottlenecking level :------------------:|:----------: $PBC2<1$ |Severe $1\le PBC2 \le 3$ |Moderate $PBC2>3$ |None

• Non-mitochondrial reads (ratio) is the percentage of non-mitochondrial read in total mapped reads. (mitochondrial reads removed).The larger value shows the better quality.

• Unique mapped reads (ratio) is the percentage of non-mitochondrial unique mapped read in total mapped reads (multi-mapped reads removed). The larger value shows the better quality.

• Duplicate removed reads (final for use) is the percentage of non-mitochondrial, unique mapped and non-duplicate reads in total mapped reads (duplicate reads removed). These reads are ready to use and storage at final. The larger value shows the better quality.

• Nucleosome free reads (<100bp) is the nucleosome free reads reads shorter than 100bp for peak calling

• Total peaks is the number of peak called by using nucleosome free reads (<100bp)

• Peaks overlaped with union DHS (ratio) is the percentage of called peak overlaped with blacklist. The larger value shows the better quality.

• Peaks overlaped with blacklist (ratio) is the percentage of called peak overlaped with blacklist. The smaller value shows the better quality.

• Fraction of reads in peaks (FRiP) is the fraction of nucleosome free reads (<100bp) in peak. The larger value shows the better quality.

Session Info

sessionInfo()

wzthu/ATACpipe documentation built on Aug. 12, 2022, 7:38 a.m.